Nanoparticle-Assisted Electron Injection

• Nanoparticle-assisted electron injection is a process where plasmonic nanoparticles generate non-equilibrium hot carriers that transfer across interfaces to adjacent acceptors.
• Studies show that nanoparticle geometry, resonance, and interfacial bonding critically influence the electron injection efficiency and mechanism in systems like Au/TiO2 and plasma accelerators.
• Experimental signatures such as transient absorption, current-voltage asymmetry, and symmetry-breaking nonlinear optical responses robustly indicate the underlying mechanisms of electron injection.

Searching arXiv for recent and foundational papers on nanoparticle-assisted electron injection to ground the article in the current literature. arXiv search results considered for this article include recent mechanistic, theoretical, and application-specific works on plasmon-assisted hot-carrier transfer, Schottky injection, adsorbate-level renormalization, nonlinear-optical readout of injected charge, and nanoparticle-triggered wakefield injection, including (Guan et al., 27 Mar 2025, Kiani et al., 2024, Lei et al., 2023, Ng et al., 2017, Govorov et al., 2013, Khurgin, 2019, Orlanducci et al., 2024, Kutrovskaya et al., 2022, Soria et al., 2022, Aniculaesei et al., 2019), and (Aniculaesei et al., 2022). Nanoparticle-assisted electron injection denotes a class of processes in which a nanoparticle enables, enhances, or localizes electron transfer into an adjacent acceptor by supplying non-equilibrium carriers, reshaping interfacial energetics, or perturbing the local field and trapping landscape. In nanoplasmonics and interfacial optoelectronics, the term most commonly refers to plasmonic or photoexcited nanoparticles that generate hot carriers and inject them into semiconductors, molecules, liquids, or low-dimensional solids. In a broader usage found in plasma acceleration, it also denotes nanoparticle-triggered trapping of electrons into a laser-driven wake. Across these contexts, the central issues are the microscopic origin of the injected electrons, the role of interfacial barriers and momentum selection, the importance of nanoparticle geometry and resonance, and the distinction between true charge transfer and purely electromagnetic enhancement (Govorov et al., 2013, Khurgin, 2019, Guan et al., 27 Mar 2025, Špádová et al., 14 Aug 2025).

1. Conceptual scope and terminological boundaries

In plasmonic metal/acceptor heterostructures, nanoparticle-assisted injection usually begins with optical excitation of a localized surface plasmon resonance, followed by non-radiative plasmon decay into non-equilibrium carriers and then transfer of a subset of those carriers across a metal/acceptor interface. This framework underlies studies of Au nanorod/TiO$_2$ Schottky systems, Au nanoparticle/carbon-bundle junctions, Au/p-GaN photocathodes with ferricyanide reduction, Au-decorated nanodiamonds emitting hydrated electrons into water, and Au/bilayer-MoS$_2$/Au-film hybrids in which injected charge is inferred from symmetry breaking (Ng et al., 2017, Kutrovskaya et al., 2022, Kiani et al., 2024, Orlanducci et al., 2024, Guan et al., 27 Mar 2025).

A narrower but foundational theoretical formulation treats photo-injection from an optically excited metal nanocrystal into a semiconductor or molecule as a confinement-controlled hot-carrier problem. In that picture, the decisive change relative to bulk metal is non-conservation of momentum in the nanocrystal, which allows the hot-carrier distribution to extend toward $E_F+\hbar\omega$ in sufficiently small nanocrystals and thereby makes over-barrier injection feasible (Govorov et al., 2013). A complementary theory of fundamental limits emphasizes that the relevant injected carriers are not a thermalized “hot-electron gas,” but primarily first-generation quasi-ballistic carriers created directly by plasmon decay, with extraction strongly limited by electron-electron scattering and interface transmission (Khurgin, 2019).

The same phrase also appears in laser wakefield acceleration, where a nanoparticle embedded in a gas target is ionized by the driver and creates a localized electrostatic perturbation that enables electron trapping into the plasma wake. In that usage, the injected electrons need not cross a solid-state interface at all; instead, the nanoparticle acts as a localized injection trigger for plasma electrons and nanoparticle-born electrons (Aniculaesei et al., 2019, Aniculaesei et al., 2022, Špádová et al., 14 Aug 2025). This broader usage suggests that “nanoparticle-assisted electron injection” is best understood as a family of nanoparticle-mediated injection processes rather than a single mechanism.

2. Microscopic mechanisms and interfacial energetics

The dominant plasmonic sequence is localized plasmon excitation, non-radiative decay into hot carriers, and interfacial transfer of carriers with sufficient energy to overcome an interfacial barrier. In the Au nanodimer@bilayer-MoS$_2$@Au-film heterostructure, the proposed direction of transfer is from the Au nanoparticle into the upper MoS$_2$ layer adjacent to the particle under 532 nm CW excitation resonant with the nanodimer plasmon (Guan et al., 27 Mar 2025). In Au nanorod/TiO$_2$, injection across a Schottky contact was traditionally framed as internal photoemission from an isotropic hot-electron population, but transient-absorption data and contact-area scaling instead supported a surface charge emission mechanism in which the injection rate is proportional to $|E_n|^2$, where $E_n$ is the electric-field component normal to the interface (Ng et al., 2017).

A key theoretical result is that nanoparticle size and confinement control whether a plasmon generates mostly near-$E_F$ carriers or a broad distribution capable of injection. For metal nanocrystals with one confined dimension, the central parameter is

$\Delta n=\frac{\omega}{q_L v_F},$

with $_2$0. In gold nanocrystals, efficient high-energy hot-electron generation occurs when $_2$1, and prism-like nanoantennas with a small width of about $_2$2–$_2$3 nm are particularly favorable when the optical field is polarized along that small dimension (Govorov et al., 2013). A separate limit theory reaches a related conclusion from transport rather than confinement: after even one electron-electron scattering event, the carrier energy is usually too diluted for efficient extraction, so injection is dominated by first-generation carriers generated by Landau damping and phonon- or defect-assisted intraband absorption (Khurgin, 2019).

Interface energetics are not set solely by barrier height. In the many-body study of $_2$4 on plasmonic Au, the quasiparticle electron-injection barrier is the LUMO-like quasiparticle energy relative to the metal Fermi level, $_2$5, and its spatial variation arises from competition between substrate plasmonic polarization and molecule–metal hybridization (Lei et al., 2023). On Au(111) nanoparticle facets, the lowest barriers occur near the facet center rather than the edge, even though edge fields are stronger. This directly contradicts the widespread assumption that electromagnetic hot spots automatically define the most favorable injection sites (Lei et al., 2023).

In electrochemical nanoparticle-assisted injection, the transfer channel may deviate from a purely outer-sphere tunneling picture. In Au/p-GaN photocathodes reducing ferricyanide, the data and modeling support coexistence of high-energy outer-sphere transfer and low-energy inner-sphere transfer mediated by CN-ligand adsorption on Au, with the ferricyanide LUMO near $_2$6 above the Fermi level (Kiani et al., 2024). In that system, the interband regime of Au remains productive because abundant low-energy electrons can transfer by the inner-sphere pathway rather than being lost as sub-threshold carriers (Kiani et al., 2024).

A distinct but related case is adsorbate-sensitized semiconductor nanoparticles. For Alizarin on TiO$_2$7, the injection mechanism depends on adsorption mode: tridented adsorption gives an indirect mechanism, whereas bidented and chelated adsorption give direct interfacial charge transfer (Soria et al., 2022). This suggests that nanoparticle-assisted injection can be controlled not only by the nanoparticle acting as donor, but also by nanoparticle morphology determining which interfacial bonding motif is available.

3. Experimental signatures and readout methodologies

Transient absorption, current–voltage asymmetry, photoelectrochemical quantum efficiency, and nonlinear optical selection rules are among the main observables used to infer nanoparticle-assisted injection. In Au nanorod/TiO$_2$8, visible-pump/near-IR-probe transient absorption showed a positive near-IR signal for Au/TiO$_2$9 but not for Au on glass, and partially embedded rods exhibited a $E_F+\hbar\omega$0 larger relative injection quantum yield than non-embedded rods. Because this enhancement was much smaller than the geometric increase in contact area but close to the $E_F+\hbar\omega$1 prediction of a field-localized surface-emission model, the data argued against a homogeneous isotropic hot-electron gas picture (Ng et al., 2017).

In Au nanoparticle/carbon-wire-bundle composites probed by STM, 532 nm illumination near the plasmon resonance produced a pronounced bias asymmetry in the $E_F+\hbar\omega$2-$E_F+\hbar\omega$3 characteristics over carbon-bundle regions, while current collected over bare Au nanoparticle regions remained symmetric. The junction resistance decreased with increasing optical intensity, and the authors associated the effect with hot carriers generated in Au nanoparticles and transferred into the carbon bundles (Kutrovskaya et al., 2022). Their formal transport analysis used the Landauer-Büttiker expression

$E_F+\hbar\omega$4

and they argued that an equilibrium hot-electron distribution could not explain the observed asymmetry (Kutrovskaya et al., 2022).

In AuNP@DND/water, ultrafast transient absorption at 720–725 nm provided a direct spectroscopic signature of hydrated electrons generated under visible excitation. The hydrated-electron signal rose on a $E_F+\hbar\omega$5 ps timescale, was suppressed by $E_F+\hbar\omega$6–$E_F+\hbar\omega$7 by NaNO$E_F+\hbar\omega$8, and was absent for pure water, bare DND, and citrate-stabilized 10 nm Au nanoparticles under the corresponding visible conditions (Orlanducci et al., 2024). The transient absorbance was defined as

$E_F+\hbar\omega$9

and the injection efficiency was estimated through Lambert–Beer analysis as $_2$0 for AuNP@DND under 500 nm excitation at $_2$1 (Orlanducci et al., 2024).

A particularly distinctive optical probe appears in bilayer MoS$_2$2. In pristine 2H bilayer MoS$_2$3, inversion symmetry makes the second-order response vanish, but injected charge from a plasmonic Au nanodimer breaks the equivalence of the two layers and turns on SHG. The second-order polarization is written as

$_2$4

and power-dependent SHG saturates above about $_2$5, consistent with a capacitor-like picture of interfacial charge accumulation (Guan et al., 27 Mar 2025). The injection model defines

$_2$6

so increasing injected charge raises the barrier and reduces further injection efficiency (Guan et al., 27 Mar 2025). Polarization-resolved SHG fitting gave $_2$7, interpreted as a quantitative measure of the out-of-plane symmetry breaking induced by injected electrons (Guan et al., 27 Mar 2025).

These studies also clarify a methodological point: many experiments measure not the transfer event itself but a downstream consequence—transient absorption in an acceptor, current asymmetry, quantum efficiency, or a symmetry-sensitive nonlinear optical tensor. This suggests that the evidentiary chain in nanoparticle-assisted injection is often indirect, and that mechanism assignments depend strongly on the specificity of the chosen observable.

4. Spatial, structural, and chemical determinants

Nanoparticle-assisted injection is highly position dependent. For $_2$8 adsorbed on a truncated-octahedral $_2$9 nanoparticle, the LUMO-like electron-injection barrier varies by as much as $_2$0 between center and edge positions, while the HOMO-like hole-injection barrier varies by up to $_2$1 (Lei et al., 2023). The most favorable electron injection sites are $_2$2 and $_2$3, each with a barrier of $_2$4 above the Fermi level, and the most favorable hole injection sites are $_2$5, $_2$6, and $_2$7, with the HOMO-like level $_2$8 below the Fermi level (Lei et al., 2023). The spatial pattern arises from an intertwined competition between purely plasmonic couplings and molecule–substrate hybridization rather than from field enhancement alone (Lei et al., 2023).

Geometry also enters through field localization. Finite-element analysis of nano-spiked plasmonic cavity arrays showed that closely spaced spike pairs create hot-spot fields, strong surface charge localization, and enhanced power dissipation in the inter-tip cavity. Tip separations of 10 nm and 2 nm were examined, and decreasing proximity caused a red shift and increased energetic electron density on the surface (Banerjee et al., 2023). That work did not compute actual injected current across a Schottky interface, but it treated the hot-spot field and surface-charge distribution as precursors to photo-injection across a barrier (Banerjee et al., 2023).

Chemical bonding at the acceptor interface can be equally decisive. In the Au/p-GaN ferricyanide system, low-energy electrons are captured efficiently only because ferricyanide can engage in an inner-sphere pathway via CN-ligand-mediated coupling to Au (Kiani et al., 2024). In Alizarin/TiO$_2$9, whether the first excitation is dye-centered or directly charge-transfer in character depends on whether the molecule is tridented, bidented, or chelated (Soria et al., 2022). In AuNP@DND, the evidence favors visible excitation of $_2$0-derived surface states on nanodiamond, amplified by the nearby Au plasmon, rather than direct photoemission from Au itself (Orlanducci et al., 2024).

These results help separate several distinct design variables that are often conflated. Electromagnetic enhancement controls local excitation strength; interfacial hybridization controls quasiparticle alignment and orbital mixing; adsorption geometry controls whether transfer is direct or indirect; and nanoscale morphology controls which of these regimes is even accessible. A plausible implication is that optimal injection sites need not coincide with the strongest optical hot spots.

5. Material platforms and application domains

The experimental and theoretical literature spans several materially distinct platform classes.

Platform	Injection role of nanoparticle	Acceptor or target
Au nanorod/TiO$_2$1	Plasmonic hot-electron source	TiO$_2$2 semiconductor (Ng et al., 2017)
Au nanoparticle/carbon bundles	Hot-carrier donor and contact	LLCC bundle transport medium (Kutrovskaya et al., 2022)
AuNP@DND in water	Visible absorber and near-field sensitizer	DND/water interface, hydrated electrons (Orlanducci et al., 2024)
Au/bilayer-MoS$_2$3/Au-film	Plasmonic injector and symmetry-breaking source	Bilayer MoS$_2$4 (Guan et al., 27 Mar 2025)
Au nanoparticle with adsorbed $_2$5	Plasmonic substrate controlling QP barriers	Molecular adsorbate (Lei et al., 2023)
Al nanoparticle in LWFA	Localized plasma injection trigger	Plasma wake (Aniculaesei et al., 2019)

In light-energy conversion, photodetection, and photocatalysis, the standard platform is a plasmonic metal nanoparticle attached to a semiconductor or molecule (Ng et al., 2017, Govorov et al., 2013). In photocathodes such as Au/p-GaN, the nanoparticle participates in a dual-interface architecture: holes are collected by the semiconductor, while electrons are injected into a redox species in solution (Kiani et al., 2024). In aqueous reductive chemistry, Au-decorated nanodiamonds offer a route to visible-light generation of hydrated electrons, extending diamond’s electron-emission chemistry beyond the UV regime (Orlanducci et al., 2024). In 2D materials, nanoparticle-assisted injection can function less as a charge-harvesting step than as a symmetry-control tool, as shown by hot-electron-induced SHG in bilayer MoS$_2$6 (Guan et al., 27 Mar 2025).

A different application family appears in wakefield acceleration. In the first proof-of-principle experiment, aluminium nanoparticles generated by laser ablation inside a helium nozzle triggered injection into a nonlinear wake and improved beam quality relative to self-injection, with peak energies up to 338 MeV, relative energy spread down to 4.7%, and vertical divergence down to 5.9 mrad (Aniculaesei et al., 2019). In a 10 cm nanoparticle-assisted hybrid wakefield accelerator, aluminum nanoparticles mixed into helium enabled bunches of $_2$7 with 340 pC, 3.4 GeV RMS convolved energy spread, and 0.9 mrad RMS divergence (Aniculaesei et al., 2022). A later simulation study treated nanoparticle material and size as control parameters and found that beam charge could be tuned from 10 to 600 pC, with saturation behavior determined by nanoparticle field strength and total atom count (Špádová et al., 14 Aug 2025). This usage is physically distinct from interfacial hot-carrier transfer, but it preserves the central idea that a nanoparticle can localize and control injection.

The literature also includes an internal, non-interfacial meaning of photo-injection. In high-index silicon nanoparticles with magnetic Mie resonances, femtosecond irradiation generates dense internal electron-hole plasma above $_2$8, around $_2$9 in the switching regime, which transiently changes the nanoparticle dielectric function and tunes its optical response (Makarov et al., 2015). That work does not involve electron transfer into another medium, but it shows that “injection” can also denote internal photocarrier generation inside a nanoparticle.

6. Limits, misconceptions, and unresolved issues

Several recurring misconceptions are explicitly challenged by the cited work. The first is that nanoparticle-assisted injection can be predicted from optical field enhancement alone. The $|E_n|^2$0/Au many-body study showed that the lowest charge-injection barriers do not necessarily occur at electromagnetic hot spots, because hybridization can outweigh edge-enhanced plasmonic screening (Lei et al., 2023). The second is that plasmon decay produces a homogeneous hot-electron reservoir whose isotropic flux determines injection. Au nanorod/TiO$|E_n|^2$1 instead supported a surface-emission picture tied to the local normal field (Ng et al., 2017). The third is that hot-carrier populations can be treated by a single temperature. A theory of fundamental limits argued that under realistic illumination a nanoparticle almost never hosts more than one plasmon quantum at a time and that the nonequilibrium distribution cannot be described by a single $|E_n|^2$2 (Khurgin, 2019).

The strongest theoretical limits concern carrier generation, first-collision survival, and interface transmission. For smooth interfaces, momentum conservation creates a severe escape-cone restriction, while for rough interfaces the ultimate limit shifts to the density-of-states mismatch between the metal and the accepting material (Khurgin, 2019). Near threshold, the classical Fowler law,

$|E_n|^2$3

reappears in nanocrystal injection theory, but the microscopic interpretation differs because quantum confinement and momentum relaxation in the nanoparticle determine whether a significant high-energy tail exists at all (Govorov et al., 2013).

Many experimental interpretations remain indirect. The bilayer-MoS$|E_n|^2$4 SHG study did not directly image transferred electrons or resolve an ultrafast injection timescale, and it did not fully eliminate thermal contributions by dedicated thermometry (Guan et al., 27 Mar 2025). The Au nanorod/TiO$|E_n|^2$5 study established a mechanistic inconsistency with the isotropic hot-electron model, but absolute injection quantum yield was not measured (Ng et al., 2017). The AuNP@DND work favored plasmonic enhancement of DND surface-state excitation, yet explicitly acknowledged that hot-electron transfer from Au to DND, chemical enhancement, and resonant energy transfer could not be excluded a priori (Orlanducci et al., 2024). The Au/p-GaN ferricyanide study supported inner-sphere transfer but left the operando interfacial structure to be clarified further (Kiani et al., 2024).

The most stable synthesis of the field is therefore mechanistic rather than universal. Nanoparticles assist electron injection by concentrating optical energy, generating non-equilibrium carriers, reshaping local barriers, and creating localized geometries or chemical motifs unavailable in extended systems. Whether this assistance yields useful transfer depends on at least five coupled factors: carrier generation channel, confinement-controlled carrier energy distribution, first-collision transport length, interface transmission rules, and chemical or structural specificity of the acceptor. This suggests that future progress will come less from maximizing any single quantity and more from co-engineering morphology, resonance, interfacial bonding, and acceptor density of states in a single design framework (Govorov et al., 2013, Khurgin, 2019, Lei et al., 2023, Kiani et al., 2024).